Environmental Engineering Reference
In-Depth Information
practical engineering applications other consid-
erations, apart from purely scientific ones, come
into play. Thus the quantity and quality of data can
be a very important factor in determining the type
of model used. In fact, the quality of the data is one
of the main sources of uncertainty in modelling.
Even given a perfect model, if inaccurate data are
used as input, the forecast will be affected. In
situations where the available information is of
poor quality, a model that gives qualitative fore-
casts may actually be of more practical assistance
in informing a decision, than an inaccurate fore-
cast produced by a sophisticated model.
One promising area of development is ensemble
modelling. This involves running models many
times over, with slightly different but realistic
initial conditions, to create a collection (or ensem-
ble) of possible future outcomes. The ensemble is
considered to be a sample of the population of all
possible future outcomes. The sample average and
variance can be computed directly to provide an
indication of the likely outcome and the expected
variation about this. This type of approach to han-
dlinguncertainty is currently the subject of intense
research activity. In the coastal literature there are
some examples of this type of approach, using the
simpler empirical or hybrid models (Dong and
Chen 1999; Reeve and Spivack 2004), and data-
driven models (Reeve et al. 2008). These can pro-
vide useful indications of shoreline behaviour and
sensitivity. The development and implementation
of an ensemble modelling approach using detailed
process models has yet to accomplished, and is the
subject of current research.
The concept of balancing risk and benefit is
central to engineering. It is rarely up to the engi-
neer alone to decide this balance. Nevertheless, it
is often the engineer who has the important role in
quantifying, as clearly as possible, the facts that
inform financial and political decision-making.
Designing flood defence infrastructure in the face
of uncertainty has been a primary hallmark of
engineering design for many years. Cardinal ele-
ments of probability theory have been used exten-
sively but are frequently well camouflaged in
what, on the surface, appear to be essentially
deterministic approaches to design.
erosion takes place there is a catch-up, which
would be a straight line up fromthe zero erosion
to meet the erosion profile after a set period of
time defined by the User.
Scenario 3: The last scenario considers the effect
on the erosionprofile if the defence hadnot been
in place for a certain period of time. Thus the
erosion profile is shifted to the left to assume
this would have been the erosion had the de-
fence not been in place (the User can specify
how old the defence currently is). After this
process is carried out the sheet works out a
catch-up (in a similar way to Scenario 2), to
construct the given erosion profile.
The methodology provides the User with the
ability to investigate the answers to two particular
types of question:
1 What is the probability of erosion for an asset
located at a given distance X?
2 What is the probability of erosion for a given
time in the future?
Further information on the methodology
and spreadsheet can be found in Defra (2006) and
Pedrozo-Acu ˜ a et al. (2008).
This example demonstrates how 'ownership'
of some of the uncertainty can be passed back to
the user of the technique - thereby allowing en-
gineering experience to be included directly into
the process. The visual interface has also proved to
be a powerful way to communicate the nature of
the risk to non-specialists.
Concluding Remarks
In this chapter some of the uncertainties associ-
ated with modelling coastal processes have been
discussed. Methods for identifying, quantifying
and to some extent mitigating uncertainty in
coastal modelling have been illustrated through
a set of case studies. These examples cover appli-
cations of modelling tidal flows, and statistical
and dynamical process prediction of beach profile
change as well as cliff erosion.
The different types of model that are available
for
forecasting have also been discussed.
In
Search WWH ::




Custom Search